On Tue, 20 Sep 2005, Vlad Drukker wrote:
There is a bug in linux bridge code, causing TSO not to work. MTU oversize packets silently dropped in br_dev_queue_push_xmit() Locally generated TSO packets are bigger than MTU (surprise, eh) The patch below, works for me. Consider to apply. David, is it OK to drop packets silently?
The Bridge standard mandates that oversized packets should be dropped in the bridge, but it says nothing about locally generated oversized packets..
On a related note it has been discussed a few times if this really is a desireable feature of the bridge code or not, but the fact that standards say it should when attempting to forward a frame between interfaces of different MTU weights pretty much.
This drop condition isn't recorded in any of interface statistics. Should it be?
Probably not. It's mainly a forwarding drop condition when you bridge between interfaces with different MTU, just that the bridge code didn't consider that the local stack may also hand it packets larger than the device MTU..
Regards Henrik - : send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
